PROPOSED METHOD · RSI v0.1

Representation Sensitivity Index

How strongly does a bounded change in representation propagate into the knowledge architecture built above it?

13.8
experimental RSI v0.1 / 100-scale comparative value
Scientific boundary: RSI v0.1 is a proposed experimental comparison metric, not a validated universal index. It measures structural sensitivity under a stated perturbation; it does not measure truth, importance or causality.

Layer-by-layer sensitivity

LayerChangedReference populationRate
Record exposure240223110.79%
Classification edits244223110.97%
Relationship turnover58064279.02%
Journey sensitivity3918521.08%
PROPAGATION

2.42

typed-route membership changes per changed record.

PROPAGATION

0.163

changed aggregate Journey signatures per changed record.

Formula v0.1

RSI = 100 × √(Relationship Turnover Rate × Journey Sensitivity Rate)

The geometric mean prevents either downstream layer from completely dominating the composite. Future versions can add Signature Journey and Capability-candidate sensitivity after those promotion rules are made reproducible.

Why this matters

The same experiment can be repeated across CIDM 2,231, Oxford Core 5000, Chinese 3500 and the reference population. RSI then becomes a common experimental language for asking how stable a knowledge architecture is to changes in its representations.

← Impact Simulation · RSI data → · Colleague Beta →